ON THE MAGNETIC DISTURBANCES. 491 



amounts of deflection were also divided into " deflections 

 on either side of the normal," and treated in the same 

 way. The variation in relative frequency and amount 

 in the respective periods being the object of this inves- 

 tigation, the absolute numbers and values of the sepa- 

 rated observations would have no other importance 

 than as evidencing that they were sufficiently large to 

 justify the deduction of the relative proportions. The 

 numbers and values would, of course, be larger or smaller 

 according as the amount of deflection taken as the mea- 

 sure of a disturbed observation were lower or. higher ; 

 but it was soon found on trial, that there were always 

 certain moderate limits within which this amount might 

 be taken indifferently without producing any material 

 alteration in the relative proportions. At stations where 

 the energy of the disturbing force was great in com- 

 parison with the deflections produced by the ordinary 

 and regular variations, a larger body of disturbed ob- 

 servations could be safely separated: but in no case 

 could the separated values be regarded as furnishing a 

 measure of the whole effects of the occasional disturb- 

 ances ; inasmuch as, until some other criterion than that 

 of magnitude is known, by which an effect of the par- 

 ticular causes now under notice can be discriminated, 

 there must always be supposed to be minor effects of the 

 same causes remaining amongst the unseparated body. 



By pursuing this method of investigation with the 

 observations that had been received from the observato- 

 ries of Toronto, St. Helena, and Hobarton (stations widely 

 distant from each other on the globe), the same conclu- 

 sions were obtained from each, viz. that the disturb- 

 ances are governed, in the frequency of their occurrence, 



